Enterprises rethink cloud plans as AI workloads strain infrastructure

Information Services Group said more than 80% of enterprises are reassessing their cloud strategies to support AI workloads, with no transaction value, budget impact or company-level spending disclosed. The finding points to a practical shift for cloud buyers and vendors: AI deployments are forcing enterprises to reconsider where compute, data and controls should sit, rather than defaulting to a single public-cloud posture.

ISG’s Wednesday report said enterprises are moving toward hybrid approaches that can combine public cloud, private cloud, colocation, edge infrastructure and sovereign environments. The advisory firm tied the shift to AI compute demand, resilience goals, cost pressure, performance requirements and the need for more consistent operations across infrastructure types.

Anay Nawathe, ISG’s cloud delivery lead for the Americas, said in a statement that enterprises are redesigning operating models for flexibility and cost discipline as AI requirements grow. He framed hybrid cloud less as a placement decision for workloads and more as a way for enterprises to keep control across a broader set of environments.

AI is changing the cloud procurement brief

ISG said AI expansion is a central reason enterprises are changing cloud plans. The report cited investments in GPU-enabled architectures, distributed data platforms and hybrid AI operating models, alongside adoption of AI platforms.

That is a more specific infrastructure agenda than the usual AI procurement language. GPU-heavy deployments raise different questions than general application migration: capacity availability, data movement, latency, recovery, compliance and the ability to track costs across multiple execution venues. ISG said enterprises are looking for unified platforms that bring observability, automation and financial management into one operating framework.

The report also identified performance management and cost management as priorities for enterprises pursuing hybrid cloud environments. That tracks with a broader buyer concern around AI economics, where experimentation can move quickly from pilots to materially higher infrastructure bills. ISG said rising use of GPU-intensive infrastructure is increasing demand for cost transparency and optimization across private cloud, Kubernetes, edge and sovereign cloud environments.

Cloud vendors chase AI infrastructure demand

Cloud providers and infrastructure specialists are adjusting their offerings as enterprise AI demand grows. Gartner has projected that neocloud providers, which sell AI-optimized infrastructure, will capture 20% of a $267 billion AI cloud market by 2030.

Large technology companies are also adding capacity and packaging around AI compute. Apple opened access to private cloud compute in June, while Blackstone and Google introduced a compute-as-a-service offering in May, according to the companies.

Sovereignty is another driver in ISG’s report. The firm said enterprises want more control over infrastructure, data residency and AI governance as they plan longer-term AI programs. That gives regional, private and hybrid infrastructure providers a clearer opening, especially with regulated buyers that cannot treat AI workloads as standard cloud migrations.

Shashank Rajmane, an ISG principal analyst and lead author of the report, said enterprises increasingly expect providers to combine operations, cyber recovery and cost management. The bar for vendors is therefore shifting from selling more capacity to proving they can help customers run fragmented infrastructure without losing visibility, continuity or cost control.

This story draws on original reporting from CIO Dive.